CHANGES.txt

0.9.2 - 2013-05-13
==================
Fix Alex's tests, aka "." (Kyle Wilcox)
adding tests for fluid dataset operations (Alexander Crosby)
seemingly working fluid api on dataset objects (Alexander Crosby)
working fluid api on datasets, needs some thorough testing (Alexander Crosby)
sort of working issue #55, chaining is not working (Alexander Crosby)
syntax error fix, just starting writing in engligh instead of python (Alexander Crosby)
attempted to make the dimension classes allow and ignore np.nan values in their arrays, so that nearest methods and bounding methods still work (Alexander Crosby)
first stab a dataset methods for rgrid (Alexander Crosby)
stubbing out new dataset modifying methods (Alexander Crosby)
adding NotImplementedErrors to Dataset base class in cdm module (Alexander Crosby)
typo, issue #54 should be fixed (Alexander Crosby)
more issue #54 fixes and testing (Alexander Crosby)
testing bbox (Alexander Crosby)
Merge branch 'master' of github.com:asascience-open/paegan (Alexander Crosby)
fix for issue #54 (Alexander Crosby)
Bug in Dataset's gettimebounds if querying by unit (Kyle Wilcox)
add the octant way of averaging the ROMS u / v vectors. (Kyle Wilcox)
Update the Diel test to be more obvious (Kyle Wilcox)
Particles should be timezone aware. (Kyle Wilcox)
Add timezone support to Timevar (similar to matplotlib) (Kyle Wilcox)
Add Diel cycles model run for testing (Kyle Wilcox)
Don't try to pickle the DataController after closing the NetCDF file (Kyle Wilcox)
Pass the timevar tests by not assuming the data is in UTC (but it is) (Kyle Wilcox)
Add back in the cPickle import (Kyle Wilcox)
cleaning up, removing traces of jd on timevarobject, use datenum property or get_datenum method (Alexander Crosby)
changes to timevar and parallel_manager to remove pylab/matplotlib dependency, now relying on netCDF4 conversion utilities to convert to a standard paegan origin datenum for processing and comparison (Alexander Crosby)
Little cleanup of the Cache creation code. (Kyle Wilcox)
Improve the model tests a bit (Kyle Wilcox)
So we should probably have a WORKING conversion process in Depthvar. (Kyle Wilcox)
Missed a parameter (Kyle Wilcox)
Decrement the read_count on a zombie... if it was reading. (Kyle Wilcox)
Less verbose logging when unable to find shoreline intersection (Kyle Wilcox)
Don't release locks if they can't be (Kyle Wilcox)
Fix bug that was setting the lock object instead of its value. (Kyle Wilcox)
Don't error on self.local in the finally block.  Load the dataset first. (Kyle Wilcox)
Have Diel check the previous day for behaviors as well as the present day (Kyle Wilcox)
Fix suncycles to always return same day. (Kyle Wilcox)
Add strange units to depthvar for PWS datasets (Kyle Wilcox)
Store which process PID has each lock to release on zombification (Kyle Wilcox)
Make TIME in the NetCDF output an int (Kyle Wilcox)
adding some logic to ensure rgrid datasets are producted in CommonDataset.open if the axes have the same length, but cdm_data_type attribute is set to grid (Alexander Crosby)
fix for negative indexes in asastochastic count_settlement (Alexander Crosby)
call task_done when the zombie is caught.  Decrement the n_run value too. (Kyle Wilcox)
Typo in zombie catching string substitution (Kyle Wilcox)
Put a zombied process on the result queue to handle it like the others (Kyle Wilcox)
Manually kill hanging processes, increase process check timeout. (Kyle Wilcox)
Exitcode 0 processes are zombie proof.  Don't start too many Consumers. (Kyle Wilcox)
Only log a retrieved result from a zombie or completed particle (Kyle Wilcox)
Remove dead procs from the procs list. (Kyle Wilcox)
Attempt to handle zombie processes (Kyle Wilcox)
offset by accident in asastoch for counting settlement (Alexander Crosby)
working asastochastic for counting settlement (Alexander Crosby)
paegan __init__ try catch logging crap (Alexander Crosby)
changes to asastochastic to allow for more flexibility so i can get my job done@@@git pullgit pullgit pull (Alexander Crosby)
Merge branch 'master' of github.com:asascience-open/paegan (Alexander Crosby)
. (Alexander Crosby)
"ugh" or ".", whichever Alex prefers this morning. (Kyle Wilcox)
single point shorelines need to calculate a linestring for indexing (Kyle Wilcox)
Loop to find a point offshore to test reverse with.  Try 12 times reverse. (Kyle Wilcox)
Add ability to test a single point with a shoreline. (Kyle Wilcox)
Add some better progress logging (Kyle Wilcox)
Set the progress queue to length 1 in EasyLogger (Kyle Wilcox)
Return progress as float, so they don't all round to 0. (Kyle Wilcox)
Add a filter to the ProgressHandler to only show logging.PROGRESS items (Kyle Wilcox)
Add ProgressHandler to the EasyLogger (Kyle Wilcox)
isinstance != isinsatnce (Kyle Wilcox)
Hey, buddy, add a colon after else.  Or else. (Kyle Wilcox)
Add progress logging. Pollutes the paegan __init__.py file to add log level (Kyle Wilcox)
Bathy should ignore time and leave it up to the calling function. (Kyle Wilcox)
Shoreline 'reverse' should half distance backwards until it is in water. (Kyle Wilcox)
Implement a "reverse" bathymetry reaction (and use it). (Kyle Wilcox)
Make reverse_distance configurable in the ModelController. (Kyle Wilcox)
Add "Age" as a float column to the shapefile (Kyle Wilcox)
Change the reverse distance to 500m. (Kyle Wilcox)
Don't bounce off shoreline further than you moved to get there. (Kyle Wilcox)
More logging cleanup (Kyle Wilcox)
Forget about saving progress, it breaks other things. (Kyle Wilcox)
ForceParticle can't be serialized, so return the Particle instead (Kyle Wilcox)
Don't try to load the bathymetry if it isn't available (Kyle Wilcox)
Set up tests to use new scratch area for bathy/shore/cache/output (Kyle Wilcox)
Require a bathy file (no default) (Kyle Wilcox)
Ease up on the logging (Kyle Wilcox)
Make sure to open shoreline as read only (Kyle Wilcox)
Change logging to better suit tracing a particle's life (Kyle Wilcox)
Get rid of the old shapefile library in favor of Fiona (GDAL direct) (Kyle Wilcox)
Add a Pickle export option (Kyle Wilcox)
Reduce logfile clutter (Kyle Wilcox)
Return a tuple from the Consumer, to track codes and objects. (Kyle Wilcox)
Retry a failed DAP connection in the DataController every 30 seconds. (Kyle Wilcox)
Simplify the Diel get_time method to always check for loc4d parameter (Kyle Wilcox)
Catch DAP errors before model starts (Kyle Wilcox)
Fix Diel tests (Kyle Wilcox)
Get Hokey with it!  Pickle the Timevar to disk and load it in the Particles (Kyle Wilcox)
Update Shapely and pytz (Kyle Wilcox)
typo in requirements (Kyle Wilcox)
Settlement test needs to be 24 hours long (Kyle Wilcox)
Ignore bathy when trying to settle a particle. (Kyle Wilcox)
Cleanup (Kyle Wilcox)
Push particles 100m off shore on reaction. (Kyle Wilcox)
Settlement should be inclusive.  Add have better logging. (Kyle Wilcox)
Add a new mulitpart shoreline (Kyle Wilcox)
DIEL: Don't overshoot desired depth range.  This adds larva intelligence! (Kyle Wilcox)
Fix timezone missteps in SunCycle calculations.  this may still break in some parts of the world. (Kyle Wilcox)
Add the Nick tests (Kyle Wilcox)
Add 24 hours to the SunCycle's calculated SUNSET value.  TODO for sure. (Kyle Wilcox)
Warning if the shapefile has strange geometries. (Kyle Wilcox)
Remove old timing code (Kyle Wilcox)
simpler calculation since particle_time is always >= (Kyle Wilcox)
Add Nick's behaviour file (Pacific Herring) (Kyle Wilcox)
Change behavior files to actually see Diel migration in some tests (Kyle Wilcox)
Catch a bad geometry in a shapefile (Kyle Wilcox)
Correctly figure out which active diel to use (Kyle Wilcox)
Diel should return the current day's time when querying for closest (Kyle Wilcox)
Improve logging output (Kyle Wilcox)
Cleanup (Kyle Wilcox)
Actually return the correct calculated_vss from a Capability. (Kyle Wilcox)
settlement working with compute_probability in asastochastic (Alexander)
. (Alexander)
adding some stubs for asastochastic (Alexander)
Rewrite CommonDataset into a staticmethod .open() (Kyle Wilcox)
close netcdf files in tests (Kyle Wilcox)
Update netCDF4 module (Kyle Wilcox)
Attempt to close the dataset so the process doesn't hang in production (Kyle Wilcox)
Remove logger from the CDM. (Kyle Wilcox)
Only hit bathymetry if it is below sea level (Kyle Wilcox)
Add shoreline_path to ModelController constructor. clean up tests. (Kyle Wilcox)
Fix time tests (Kyle Wilcox)
Remove old log message (Kyle Wilcox)
If no data is pulled from forcing, use previous values for transport. (Kyle Wilcox)
clean up .gitignore and netcdf files in repo (Kyle Wilcox)
Allow the shoreline dataset to be configurable.  Set spatialbuffer to 0.25. (Kyle Wilcox)
Default the horiz_chunk to 5 (Kyle Wilcox)
Change locking procedure (Kyle Wilcox)
Allow the reverse_distance to be configrable in "react" function (Kyle Wilcox)
Don't log the CDM Dataset anymore (Kyle Wilcox)
Shorelines should always react at zero depth (Kyle Wilcox)
Speed up the normalized_locations function (Kyle Wilcox)
Implement the shapefile output using Fiona and GDAL.  Speedup plz. (Kyle Wilcox)
Change the Consumer flow. (Kyle Wilcox)
Add some verbose logging to the CDM class to debug runtime errors. (Kyle Wilcox)
Fix tests that expect an exception from the transport model (Kyle Wilcox)
Add a release script so I don't always forget and have to Google it. (Kyle Wilcox)
Prepare for 0.9.1 release (Kyle Wilcox)
Do as I say, not as I do.  Huge commit. (Kyle Wilcox)
Output working with conditional variables (Kyle Wilcox)
Add more output to shapefile and netcdf file (Kyle Wilcox)
Particles should track U/V/W from hydro model and output into shapefile (Kyle Wilcox)
Behavior library passes in positive down, switch to negative down. (Kyle Wilcox)
Fix Bathymetry / Settlement.  Add a DeadLifeStage and the killing of fish. (Kyle Wilcox)
A lifestage with just settlement does not need a capability configured. (Kyle Wilcox)
Larva particles actually grow now.  fixes #40 (Kyle Wilcox)
Clean up the model controller tests (Kyle Wilcox)
Clean up the logger to use StopIteration to close the queue (Kyle Wilcox)
dave (Kyle Wilcox)
Catch a MemoryError when the queue is closed externally (Kyle Wilcox)
Externalize the logging queue (Kyle Wilcox)
Join and then close the queue. break out of loop on all errors (Kyle Wilcox)
Reverting a bunch the old fashioned way (Kyle Wilcox)
Close the queue after the handler (Kyle Wilcox)
Don't process the flag (Kyle Wilcox)
Use -1 as a flag to stop thread loop (Kyle Wilcox)
Send a final message to break out of loop (Kyle Wilcox)
Improve logging (Kyle Wilcox)
this is a ridiculous debugging cycle (Kyle Wilcox)
Close queue after handler (Kyle Wilcox)
Close the logging queue on close (Kyle Wilcox)
Raise an error when no particles are actually run rather than success (Kyle Wilcox)
Remove Manager so it always shuts down (Kyle Wilcox)
Catch more errors, log information, and please be reliable. (Kyle Wilcox)
Add more logging to DataController (Kyle Wilcox)
Return from a run (Kyle Wilcox)
Remove old commented out line (Kyle Wilcox)
random shouldn't have a parameter (Kyle Wilcox)
Cache is now an absolute path (Kyle Wilcox)
Move logger to beginning of run (Kyle Wilcox)
Allow a parameter so the model run does not delete the cache file. (Kyle Wilcox)
Move random filename to AsaRandom, add tests (Kyle Wilcox)
Remove as much matplotlib as possible (Kyle Wilcox)
Allow subpackages to share 'paegan' namespace (Kyle Wilcox)Fix Alex's tests, aka "." (Kyle Wilcox)
adding tests for fluid dataset operations (Alexander Crosby)
seemingly working fluid api on dataset objects (Alexander Crosby)
working fluid api on datasets, needs some thorough testing (Alexander Crosby)
sort of working issue #55, chaining is not working (Alexander Crosby)
syntax error fix, just starting writing in engligh instead of python (Alexander Crosby)
attempted to make the dimension classes allow and ignore np.nan values in their arrays, so that nearest methods and bounding methods still work (Alexander Crosby)
first stab a dataset methods for rgrid (Alexander Crosby)
stubbing out new dataset modifying methods (Alexander Crosby)
adding NotImplementedErrors to Dataset base class in cdm module (Alexander Crosby)
typo, issue #54 should be fixed (Alexander Crosby)
more issue #54 fixes and testing (Alexander Crosby)
testing bbox (Alexander Crosby)
Merge branch 'master' of github.com:asascience-open/paegan (Alexander Crosby)
fix for issue #54 (Alexander Crosby)
Bug in Dataset's gettimebounds if querying by unit (Kyle Wilcox)
add the octant way of averaging the ROMS u / v vectors. (Kyle Wilcox)
Update the Diel test to be more obvious (Kyle Wilcox)
Particles should be timezone aware. (Kyle Wilcox)
Add timezone support to Timevar (similar to matplotlib) (Kyle Wilcox)
Add Diel cycles model run for testing (Kyle Wilcox)
Don't try to pickle the DataController after closing the NetCDF file (Kyle Wilcox)
Pass the timevar tests by not assuming the data is in UTC (but it is) (Kyle Wilcox)
Add back in the cPickle import (Kyle Wilcox)
cleaning up, removing traces of jd on timevarobject, use datenum property or get_datenum method (Alexander Crosby)
changes to timevar and parallel_manager to remove pylab/matplotlib dependency, now relying on netCDF4 conversion utilities to convert to a standard paegan origin datenum for processing and comparison (Alexander Crosby)
Little cleanup of the Cache creation code. (Kyle Wilcox)
Improve the model tests a bit (Kyle Wilcox)
So we should probably have a WORKING conversion process in Depthvar. (Kyle Wilcox)
Missed a parameter (Kyle Wilcox)
Decrement the read_count on a zombie... if it was reading. (Kyle Wilcox)
Less verbose logging when unable to find shoreline intersection (Kyle Wilcox)
Don't release locks if they can't be (Kyle Wilcox)
Fix bug that was setting the lock object instead of its value. (Kyle Wilcox)
Don't error on self.local in the finally block.  Load the dataset first. (Kyle Wilcox)
Have Diel check the previous day for behaviors as well as the present day (Kyle Wilcox)
Fix suncycles to always return same day. (Kyle Wilcox)
Add strange units to depthvar for PWS datasets (Kyle Wilcox)
Store which process PID has each lock to release on zombification (Kyle Wilcox)
Make TIME in the NetCDF output an int (Kyle Wilcox)
adding some logic to ensure rgrid datasets are producted in CommonDataset.open if the axes have the same length, but cdm_data_type attribute is set to grid (Alexander Crosby)
fix for negative indexes in asastochastic count_settlement (Alexander Crosby)
call task_done when the zombie is caught.  Decrement the n_run value too. (Kyle Wilcox)
Typo in zombie catching string substitution (Kyle Wilcox)
Put a zombied process on the result queue to handle it like the others (Kyle Wilcox)
Manually kill hanging processes, increase process check timeout. (Kyle Wilcox)
Exitcode 0 processes are zombie proof.  Don't start too many Consumers. (Kyle Wilcox)
Only log a retrieved result from a zombie or completed particle (Kyle Wilcox)
Remove dead procs from the procs list. (Kyle Wilcox)
Attempt to handle zombie processes (Kyle Wilcox)
offset by accident in asastoch for counting settlement (Alexander Crosby)
working asastochastic for counting settlement (Alexander Crosby)
paegan __init__ try catch logging crap (Alexander Crosby)
changes to asastochastic to allow for more flexibility so i can get my job done@@@git pullgit pullgit pull (Alexander Crosby)
Merge branch 'master' of github.com:asascience-open/paegan (Alexander Crosby)
. (Alexander Crosby)
"ugh" or ".", whichever Alex prefers this morning. (Kyle Wilcox)
single point shorelines need to calculate a linestring for indexing (Kyle Wilcox)
Loop to find a point offshore to test reverse with.  Try 12 times reverse. (Kyle Wilcox)
Add ability to test a single point with a shoreline. (Kyle Wilcox)
Add some better progress logging (Kyle Wilcox)
Set the progress queue to length 1 in EasyLogger (Kyle Wilcox)
Return progress as float, so they don't all round to 0. (Kyle Wilcox)
Add a filter to the ProgressHandler to only show logging.PROGRESS items (Kyle Wilcox)
Add ProgressHandler to the EasyLogger (Kyle Wilcox)
isinstance != isinsatnce (Kyle Wilcox)
Hey, buddy, add a colon after else.  Or else. (Kyle Wilcox)
Add progress logging. Pollutes the paegan __init__.py file to add log level (Kyle Wilcox)
Bathy should ignore time and leave it up to the calling function. (Kyle Wilcox)
Shoreline 'reverse' should half distance backwards until it is in water. (Kyle Wilcox)
Implement a "reverse" bathymetry reaction (and use it). (Kyle Wilcox)
Make reverse_distance configurable in the ModelController. (Kyle Wilcox)
Add "Age" as a float column to the shapefile (Kyle Wilcox)
Change the reverse distance to 500m. (Kyle Wilcox)
Don't bounce off shoreline further than you moved to get there. (Kyle Wilcox)
More logging cleanup (Kyle Wilcox)
Forget about saving progress, it breaks other things. (Kyle Wilcox)
ForceParticle can't be serialized, so return the Particle instead (Kyle Wilcox)
Don't try to load the bathymetry if it isn't available (Kyle Wilcox)
Set up tests to use new scratch area for bathy/shore/cache/output (Kyle Wilcox)
Require a bathy file (no default) (Kyle Wilcox)
Ease up on the logging (Kyle Wilcox)
Make sure to open shoreline as read only (Kyle Wilcox)
Change logging to better suit tracing a particle's life (Kyle Wilcox)
Get rid of the old shapefile library in favor of Fiona (GDAL direct) (Kyle Wilcox)
Add a Pickle export option (Kyle Wilcox)
Reduce logfile clutter (Kyle Wilcox)
Return a tuple from the Consumer, to track codes and objects. (Kyle Wilcox)
Retry a failed DAP connection in the DataController every 30 seconds. (Kyle Wilcox)
Simplify the Diel get_time method to always check for loc4d parameter (Kyle Wilcox)
Catch DAP errors before model starts (Kyle Wilcox)
Fix Diel tests (Kyle Wilcox)
Get Hokey with it!  Pickle the Timevar to disk and load it in the Particles (Kyle Wilcox)
Update Shapely and pytz (Kyle Wilcox)
typo in requirements (Kyle Wilcox)
Settlement test needs to be 24 hours long (Kyle Wilcox)
Ignore bathy when trying to settle a particle. (Kyle Wilcox)
Cleanup (Kyle Wilcox)
Push particles 100m off shore on reaction. (Kyle Wilcox)
Settlement should be inclusive.  Add have better logging. (Kyle Wilcox)
Add a new mulitpart shoreline (Kyle Wilcox)
DIEL: Don't overshoot desired depth range.  This adds larva intelligence! (Kyle Wilcox)
Fix timezone missteps in SunCycle calculations.  this may still break in some parts of the world. (Kyle Wilcox)
Add the Nick tests (Kyle Wilcox)
Add 24 hours to the SunCycle's calculated SUNSET value.  TODO for sure. (Kyle Wilcox)
Warning if the shapefile has strange geometries. (Kyle Wilcox)
Remove old timing code (Kyle Wilcox)
simpler calculation since particle_time is always >= (Kyle Wilcox)
Add Nick's behaviour file (Pacific Herring) (Kyle Wilcox)
Change behavior files to actually see Diel migration in some tests (Kyle Wilcox)
Catch a bad geometry in a shapefile (Kyle Wilcox)
Correctly figure out which active diel to use (Kyle Wilcox)
Diel should return the current day's time when querying for closest (Kyle Wilcox)
Improve logging output (Kyle Wilcox)
Cleanup (Kyle Wilcox)
Actually return the correct calculated_vss from a Capability. (Kyle Wilcox)
settlement working with compute_probability in asastochastic (Alexander)
. (Alexander)
adding some stubs for asastochastic (Alexander)
Rewrite CommonDataset into a staticmethod .open() (Kyle Wilcox)
close netcdf files in tests (Kyle Wilcox)
Update netCDF4 module (Kyle Wilcox)
Attempt to close the dataset so the process doesn't hang in production (Kyle Wilcox)
Remove logger from the CDM. (Kyle Wilcox)
Only hit bathymetry if it is below sea level (Kyle Wilcox)
Add shoreline_path to ModelController constructor. clean up tests. (Kyle Wilcox)
Fix time tests (Kyle Wilcox)
Remove old log message (Kyle Wilcox)
If no data is pulled from forcing, use previous values for transport. (Kyle Wilcox)
clean up .gitignore and netcdf files in repo (Kyle Wilcox)
Allow the shoreline dataset to be configurable.  Set spatialbuffer to 0.25. (Kyle Wilcox)
Default the horiz_chunk to 5 (Kyle Wilcox)
Change locking procedure (Kyle Wilcox)
Allow the reverse_distance to be configrable in "react" function (Kyle Wilcox)
Don't log the CDM Dataset anymore (Kyle Wilcox)
Shorelines should always react at zero depth (Kyle Wilcox)
Speed up the normalized_locations function (Kyle Wilcox)
Implement the shapefile output using Fiona and GDAL.  Speedup plz. (Kyle Wilcox)
Change the Consumer flow. (Kyle Wilcox)
Add some verbose logging to the CDM class to debug runtime errors. (Kyle Wilcox)
Fix tests that expect an exception from the transport model (Kyle Wilcox)
Add a release script so I don't always forget and have to Google it. (Kyle Wilcox)

0.9.1 - 2012-11-14
===================
Do as I say, not as I do.  Huge commit. (Kyle Wilcox)
Output working with conditional variables (Kyle Wilcox)
Add more output to shapefile and netcdf file (Kyle Wilcox)
Particles should track U/V/W from hydro model and output into shapefile (Kyle Wilcox)
Behavior library passes in positive down, switch to negative down. (Kyle Wilcox)
Fix Bathymetry / Settlement.  Add a DeadLifeStage and the killing of fish. (Kyle Wilcox)
A lifestage with just settlement does not need a capability configured. (Kyle Wilcox)
Larva particles actually grow now.  fixes #40 (Kyle Wilcox)
Clean up the model controller tests (Kyle Wilcox)
Clean up the logger to use StopIteration to close the queue (Kyle Wilcox)
dave (Kyle Wilcox)
Catch a MemoryError when the queue is closed externally (Kyle Wilcox)
Externalize the logging queue (Kyle Wilcox)
Join and then close the queue. break out of loop on all errors (Kyle Wilcox)
Reverting a bunch the old fashioned way (Kyle Wilcox)
Close the queue after the handler (Kyle Wilcox)
Don't process the flag (Kyle Wilcox)
Use -1 as a flag to stop thread loop (Kyle Wilcox)
Send a final message to break out of loop (Kyle Wilcox)
Improve logging (Kyle Wilcox)
this is a ridiculous debugging cycle (Kyle Wilcox)
Close queue after handler (Kyle Wilcox)
Close the logging queue on close (Kyle Wilcox)
Raise an error when no particles are actually run rather than success (Kyle Wilcox)
Remove Manager so it always shuts down (Kyle Wilcox)
Catch more errors, log information, and please be reliable. (Kyle Wilcox)
Add more logging to DataController (Kyle Wilcox)
Return from a run (Kyle Wilcox)
Remove old commented out line (Kyle Wilcox)
random shouldn't have a parameter (Kyle Wilcox)
Cache is now an absolute path (Kyle Wilcox)
Move logger to beginning of run (Kyle Wilcox)
Allow a parameter so the model run does not delete the cache file. (Kyle Wilcox)
Move random filename to AsaRandom, add tests (Kyle Wilcox)
Remove as much matplotlib as possible (Kyle Wilcox)
Allow subpackages to share 'paegan' namespace (Kyle Wilcox)

0.9.0 - 2012-10-01
===================
Improve the model_controller logging and spawning process.  fixes #33
User INFO messages in parellel_manager, since a DEBUG logger hangs
Avoid conflict in naming by importing time as timer.
Manage an EasyLogger for testing in the setUp and tearDown methods
Remove existing handlers when using the EasyLogger
Apparently setting the logger level to logger.DEBUG hangs everything.
Add test to start a run on land
Add sleeping and debug messages. Set testing logger to DEBUG by default.
Merge remote branch 'origin/larva_alex'
when failing a consumer, change get_data.value to False so the remaining particles crash
Wrap more functionality in a try/catch to avoid locking woes
Add logger when needed...
Test if we are dropped particles on land to start with.
Always close nc files and set particle_get to False
some fixes for large particles, asastochastic
working stochastic
initial commit of asastochastic.py for computing probability grids

0.8.9 - 2012-09-07
===================
fixing alex's inability to run the tests.  comments #29
Merge remote branch 'origin/larva_alex'
Implement Settlement, fixes #27
first stab at paegan issue #29
Merge remote branch 'origin/larva_alex'
Return something from getboundingpolygon
Add settlement
Add check for CPU count
changes to cdm for issue #26
Merge remote branch 'origin/larva_alex'
Fix the interp caching so we don't add spaces to the cache file
Don't subset the time_indexes until looping
Simplify the log messages when forcing particles
Fix bug when using 'nearest' time_method
Make the time_method configuration on ModelController
Get the time indexes right for forcing and saving particle locations.
Fix normalization of indexes on particle class.
changes to timevar so you can find the indices of nearest without a list or array
Move export functions to their own file and to classmethods
fixes to cdm

0.8.8 - 2012-08-24
===================
Fix the timevar tests
.
more fixes to t bounds handling in get_values in commondataset
fixing get_tind_from_bound implementation in get_values
.
added getvariableinfo on common dataset
changes to cdm to add getbbox from the commondataset (still need variable reference
same temp and salt fixes added to the netcdf output
Stick particles if no shoreline type is given
Implement a trackline output in GeoJson.  fixes #25
Normalize the length of locations/temps/salts on export
Factor the timestep calculation into a method
Don't try to save the particles returned as -1
Add the default output directory to .gitignore
Add a close method to the EasyLogger
Fix errors in model controller tests
Zip up shapefiles
Merge branch 'larva_alex'
update test with easylogger
working EasyLogger
added EasyLogger to multi_process_logging
added option to run particles with nearest OR linear interp
adding some comments, removing prints
added linear interp between timesteps for u, v, w, temp, salt
Change shapefile creation, test the output formats
added consumer messages to the logger for failed parallel tasks, also implemented a before index in cdm/timevar
changed broken changes back temporarily
Merge branch 'master' of github.com:asascience-open/paegan
Support python 2.6 by adding a monkeypatched NullHandler()

0.8.7 - 2012-08-22
===================
Prepare for 0.8.7 release
Fix typo in logging
Implement logging
Merge remote branch 'origin/larva_alex'
added output folder and formats to be specified in run method on model controller, also updated model controller __str__
catching failed tasks in the parallel consumer, failed things return None
filter out nan in nearest 4 mean calculation
Catch math.isnan potential in the behavior model
Use math.isnan to avoid forcing with nans.
Fix bug with linear growth, add test for lifestage with linear growth
Alex's mistake again
alex's mistake
Don't require unicode